Skip to content

Latest commit

 

History

History

iciar2020

Folders and files

NameName
Last commit message
Last commit date

parent directory

..
bin
bin
 
 
model_configs
model_configs
 
 
README.md
README.md
 
 
data
data
 
 
gaudio_iciar2020_after_reviewers.pdf
gaudio_iciar2020_after_reviewers.pdf
 
 
gaudio_iciar2020_before_reviewers.pdf
gaudio_iciar2020_before_reviewers.pdf
 
 
pic1.png
pic1.png
 
 
pic2.png
pic2.png
 
 
pic3.png
pic3.png
 
 
pic4.png
pic4.png
 
 
requirements.txt
requirements.txt
 
 

README.md

Code accompanying the Pixel Color Amplification paper,

A. Gaudio, A. Smailagic, A. Campilho, “Enhancement of Retinal Fundus Images via Pixel Color Amplification,” In: Karray F., Campilho A., Wang Z. (eds) Image Analysis and Recognition. ICIAR 2020. Lecture Notes in Computer Science, vol 11663. Springer, Cham (accepted)

(to be updated upon formal publication by Springer): The final authenticated publication is available online at https://doi.org/[insert DOI]

Reproducing Results

The below steps should completely reproduce results for the paper.

I tried to leave this in its "raw" state so the code is as close as possible to the state it was in when I generated results. This means it isn't cleaned up in some places. I did move many files around since generating results. I also released a simplepytorch library in order to make this reproducible.

There are three git tags identifying where the work was done.

  • iciar2020_not_reproducible - Deep network results were obtained the work using a private library that I have subsequently open sourced.
  • iciar2020 - I attempted to make the results completely reproducible.
  • iciar2020_v2 - Some code changes to address reviewer comments. Note that the paper before reviewers was 20 pages, and I was asked to cut it to 14 to length requirements. Thus, several results obtained are not published.
check out the git tag to reproduce results
$ git checkout iciar2020_v2

# current working directory should be same as this README.md

# download and unzip the datasets into ./data.  Make a directory structure like this:
 $ mkdir data
 $ ls data/{IDRiD_segmentation,RITE,arsn_qualdr}
data/arsn_qualdr

data/IDRiD_segmentation:
'1. Original Images'  '2. All Segmentation Groundtruths'   CC-BY-4.0.txt   LICENSE.txt

data/RITE:
AV_groundTruth.zip  introduction.txt  read_me.txt  test  training

# QualDR results (not used in published version of paper)
cat ./bin/qualdr/reproduce_qualdr_grading.sh

# RITE and IDRiD results  (RITE not used in the published version of paper)
cat ./bin/segmentation/reproduce.py

# Separability on IDRiD train (This goes exhaustively through all models, all images, all label types, which is probably unnecessary).
python ./bin/separability/gen_histograms.py

# plots and tables
./bin/all_analysis.sh